零膨胀联合脆弱模型在含终点事件的临床复发数据中的应用  

Application of Zero-Inflation Joint Frailty Model for Analysis of Clinical Recurrent Event Data with a Terminal Event

作　　者：郭强[1] 罗天娥[1] 于智凯 段燕[2] Guo Qiang;Luo Tiane;Yu Zhikai(Shanxi Medical University（030001）,Taiyuan)

机构地区：[1]山西医科大学卫生统计教研室,030001 [2]山西省肿瘤医院

出　　处：《中国卫生统计》2018年第5期659-662,共4页Chinese Journal of Health Statistics

基　　金：国家自然科学基金项目(81001294)

摘　　要：目的探讨零膨胀联合脆弱模型在含终点事件的临床复发数据中的应用。方法收集肺癌患者术后随访资料,构建零膨胀联合脆弱模型,采用高斯求积法进行参数估计,并与联合脆弱模型进行比较。结果零膨胀联合脆弱模型既考虑了复发事件与终点事件间的关联,又对那些未发生复发事件的"结构零"个体建立logistic模型,提高了数据拟合效果,结果解释合理,软件实现便捷。结论零膨胀联合脆弱模型适合分析临床研究中含终点事件的零膨胀复发事件数据。Objective To explore the application of zero-inflation joint frailty model for analysis of clinical recurrent e- vent data in the presence of a terminal event. Methods The follow-up data were collected from postoperative patients with lung cancer. Zero-inflation joint frailty model was fitted with the data and completed with joint frailty model. Parameters were esti- mated with Gaussian quadrature. Results Zero-inflation joint frailty model not only considers the dependent between recurrent event and terminal event, but also fits logistic model for those ＂structural zeros＂ of without any recurrent events. So, the model improves the fitted effect of the data. The results are easy to implement with software and it＇ s explain is reasonable. Conclusion The zero-inflation joint frailty model is suitable for analyzing zero-inflated recurrent event data in the presence of a terminal e- vent.

关 键 词：复发事件数据 零膨胀 终点事件 联合脆弱模型 高斯求积 

分 类 号：R195.1[医药卫生—卫生统计学]